April 16, 2021

### Bank Borrowings by Asset Managers Evidence from US Open-End Mutual Funds and Exchange-Traded Funds

#### Introduction

Open-end mutual funds and exchange-traded funds ("ETFs") engage in liquidity transformation by offering daily redemptions to investors, notwithstanding the liquidity profile of funds' underlying assets. Such liquidity transformation, combined with a first-mover advantage, could lead to investor runs and asset fire sales (Coval and Stafford 2007, Chen, Goldstein, and Jiang 2010, Goldstein, Jiang, and Ng 2017). The heavy investor redemptions experienced by many of these funds in March 2020 was a manifestation of their structural vulnerabilities.

While it is well-known that mutual funds and ETFs use various tools for liquidity risk management, including cash reserves, short-term liquid investments, and derivatives, little is known about the scope of the funds' access to and reliance on credit lines from banks. Although liquidity management in normal times mostly favor the convenience of using cash reserves and liquid derivatives such as futures, cash and derivatives give lower returns compared with returns from relatively illiquid investments and can thereby reduce investment efficiency. As such, many mutual funds open and maintain credit lines with banks to prepare for times of turmoil when it becomes extremely strenuous to rely solely on cash reserves or derivatives.

In this note, we look into investment funds' access to and usage of bank credit, based on a new dataset on credit line (and other types of loan) extension by top bank holding companies to open-end mutual funds and ETFs in the United States. We find that the aggregate amount of bank lending to open-end funds and ETFs was small and greatly fluctuated across time. Bank credit, particularly in the form of credit lines, has offered funds a flexible liquidity source from which they can draw down cash in times of excessive fund outflows, such as during the onset of the COVID-19 pandemic outbreak. In particular, funds that hold relatively more illiquid assets, such as bank loan funds, are more reliant on bank credit lines.

#### Data and Methodology

In addition, because outflows are hard to predict, multiple funds under the same fund family most often share the bank lines of credit. In the above examples, T. Rowe Price Floating Rate Fund, Inc. and T. Rowe Price Floating Rate Multi-Sector Account Portfolio shared the $800 million credit line. John Hancock Funds' multiple lines of credit were shared by 183 John Hancock funds. BNY Mellon's lines of credit from Citibank were shared by 107 funds under BNY Mellon. #### Who Borrows from Banks? Certain fund types have a higher need for bank lines of credit due to their predominantly illiquid holdings and thus a potential for unusually rapid investor redemptions. For a subset of FR Y-14 obligors for whom we successfully identified Lipper class types via string-matching, Table 2 lists the Lipper class types of mutual funds and ETFs that had the largest number of total credit lines (revolving and non-revolving) from banks. Illiquid fund types such as bank loan funds (or so-called Loan Participation funds by Lipper), high yield bond funds, real estate funds, municipal bond funds, and emerging markets funds are the largest fund classes that borrowed from banks. This observation is consistent with and extends the results from Chernenko and Sunderam (2016) that find funds with illiquid assets and flow volatility to hold larger amounts of pure cash holdings. We find that these same (illiquid) funds not only have higher levels of cash but also maintain larger amounts of loans and credit lines from banks in case that their need for liquidity exceeds their cash holdings. ##### Table 2. Top 20 Lipper Classes of Mutual Funds and ETFs by Total Number of Credit Lines from Banks, 2013:Q1-2020:Q3 Number of Credit Lines Average Loan Size ($Mil) 803 87 594 65.7 334 68.7 323 85.3 258 109 142 35.3 130 124.3 90 4.1 89 14.4 78 85.6 66 39.9 61 89.5 60 143.1 60 84.8 51 110.4 47 18.7 42 32.4 42 72.1 41 57 39 74.3

Source: Center for Research in Security Prices, FR Y-14 H1, staff calculation.

#### Bank Lending to Open-End Mutual Funds and ETFs

Figure 1(a) shows the total amount of bank lending—in the forms of term loans and committed credit lines combined—of U.S. mutual funds and ETFs from large banks, from 2013:Q1 to 2020:Q3. Over the period, the amount of large bank credit that these funds altogether had access to fluctuated in the range of $28-40 billion. The reduction in the number of unique borrowers and the average amount of loan both contributed to driving down the aggregate amount of credit extension to open-end funds and ETFs since 2015.7 As a share of the funds' total net assets, the total credit extension corresponded to 0.1-0.3% of the funds' total assets, implying that the dependence of the aggregate mutual fund industry as a whole on bank credit has stayed fairly low.8 By contrast, the economic magnitude of funds' reliance on bank credit is likely to be higher for the fund types with extremely illiquid holdings, such as those from Table 3. To learn about the differential importance of bank credit for these funds, we focus on bank loan open-end mutual funds and bank loan ETFs. Unlike the approach taken for the whole mutual fund industry, we identify these funds at the individual fund-level using a more sophisticated matching. Our final sample identifies and follows at least 40 unique bank loan funds. Figure 1(b) shows that these funds maintained, in aggregate, a bank credit amounting to$1.4-3.2 billion, or 1.2-3.4% of their total net assets during the sample period. In other words, bank credit maintained by bank loan funds was larger than that maintained by the whole industry, by a factor of around 10.9 As Anadu and Cai (2019) finds that bank loan funds held around 5% of total net assets in cash and cash-equivalent holdings in recent years, the economic magnitude of bank credit commitment is substantial from the bank loan fund's point of view. The magnitude implies that bank loan funds had an additional liquidity buffer of an average of 2% of their total assets on top of their liquidity holdings of around 5% of total assets.

#### Investor Flows and Credit Line Drawdowns

While the previous subsection looked at the overall credit extension by banks to open-end funds and ETFs, this subsection examines the actual drawdowns of the extended credit, in particular, in relationship to fund net flows. Once a credit line is drawn, the number of days that the line of credit are in use most often varies from one day to a few weeks, implying that the drawdown is mostly likely used to bridge the time gap between investor redemptions and the sales and settlement of underlying assets. Therefore, we would expect to see a negative correlation of (lagged) fund net flows and its credit line usage. 10

Figures 2(a) and (b) show how the drawdown of credit lines, as a share of total committed credit lines, fluctuates with net flows, for all mutual funds/ETFs and bank loan mutual funds/ETFs, respectively. Consistent with our expectation, we observe an overall negative relationship between them, for the whole fund industry and also for bank loan funds. Bank loan funds' highly flow-dependent reliance on bank credit is consistent with Anadu and Cai (2019)'s observation that bank loan funds increase the drawdown on credit lines in periods of stress with large outflows. It is relatively more challenging to interpret the time trend, that is, the decline in bank loan funds' drawdown of individual credit lines, because this sample of credit lines only comprises of individual lines extended to each bank loan fund and does not include lines of credit that a bank loan fund shares with other funds in the same fund family. Therefore, the true magnitude of bank credit drawdown by bank loan funds may be higher than the conservative estimate here.

#### Bank Borrowings and Credit Line Drawdowns by Open-End Funds and ETFs during the COVID-19 Outbreak

In the wake of the COVID-19 outbreak, many funds—especially those that held corporate debt and bank loans—experienced heavy investor redemptions. Table 3 shows the development in total credit extension by large banks to open-end funds and ETFs and the drawdowns before and after the pandemic outbreak. As of the end of March, mutual funds and ETFs increased their drawdown on revolving credit lines amid heavy quarterly outflows, despite a decline in banks' total credit commitment to these funds. In particular, the drawdown by bank loan funds drawdown increased by around 40%, from $190 million to$270 million, which corresponded to an increase in the percent of used credit lines from 11% to 17%. While still modest, these jumps in the credit line usage by bank loan funds were larger in magnitude than those shown in the whole sample of mutual funds and ETFs, suggesting that the drawdown in 2020:Q1 was indeed related to the liquidity management.

Although banks seem to have reduced their overall exposure to mutual funds and ETFs since the outbreak of COVID-19, the total available bank credit to bank loan funds actually increased to more than 6% of the funds' total assets as of 2020Q1, due to the shrinking size of bank loan fund assets. Since around 83% of the credit lines remains unused, this untapped liquidity amounts to at least 5% of bank loan funds' total net assets, which can be of use on top of their 5% cash and cash equivalent on average (according to Anadu and Cai (2019)). This additional buffer would help with the funds' short-term liquidity needs along with SEC's temporary borrowing relief on mutual funds that allow them to temporarily deviate from fundamental investment policies in terms of borrowing and lending and ease restrictions on interfund lending and borrowing from affiliates.

##### Table 3. Borrowings from Large Banks by U.S. Open-end Mutual Funds and ETFs around the COVID-19 Outbreak
All open-end mutual funds and ETFs Bank loan mutual funds and ETFs 2019:Q4 2020:Q1 2020:Q3 2019:Q4 29.2 28.3 27.9 1.7 1.6 1.4 3.5 3.8 3.9 0.05 0.03 0.02 32.7 32.1 31.8 1.8 1.7 1.4 0.15% 0.18% 0.14% 4.50% 6.40% 5.20% 5.2 5.5 5.1 0.19 0.27 0.16 17.90% 19.30% 18.40% 10.80% 16.60% 11.90%

Note: "Other types" include term loans and other types of credit facilities. Used (Drawn) funds are only reported for revolving credit line facilities. The sample of bank loan mutual funds in this table comprises of individual lines extended to each bank loan fund and does not include lines of credit that a bank loan fund shares with other funds in the same fund family. Therefore, the sample is likely to be smaller than the true population and so are the estimated amounts of committed and drawn credit lines by the bank loan mutual funds in the table than their true amounts. Source: Center for Research in Security Prices, Morningstar, Inc., FR Y-14 H1, staff calculation.

#### Conclusion

In this note, we look into the supervisory FR Y-14 data to examine banks' credit extension to open-end mutual funds and ETFs in the United States, and especially around the onset of COVID-19. We find that, while it is common for mutual funds and ETFs to have credit lines with banks, the overall dependence of the aggregate mutual fund industry on bank credit has remained fairly low. However, funds investing in relatively illiquid assets, such as bank loan funds, are more reliant on bank revolving credit lines that serve as a flexible source of liquidity in times of stress. Drawdown of credit lines negatively move with net fund flows, as evident in the March 2020 episode.

#### References

Anadu, K. and Cai, F., 2019. "Liquidity Transformation Risks in U.S. Bank Loan and High-Yield Mutual Funds", FEDS Notes No. 2019-08-09.

Chen, Q., Goldstein, I. and Jiang, W. 2010. "Payoff complementarities and financial fragility: Evidence from mutual fund outflows", Journal of Financial Economics, vol. 97(2), pp. 239-262.

Chernenko, S. and Sunderam, A., 2016. "Liquidity Transformation in Asset Management: Evidence from the Cash Holdings of Mutual Funds," Fisher College of Business Working Paper No. 2016-03-05, Charles A. Dice Center Working Paper No. 2016-5.

Center for Research in Security Prices. CRSP Survivor-Bias-Free US Mutual Fund Database.

Coval, J. and Stafford, E., 2007. "Asset fire sales (and purchases) in equity markets," Journal of Financial Economics, vol. 86(2), pp. 479-512.

Falato, Antonio, Ali Hortaçsu, Dan Li, and Chaehee Shin. "Fire-Sale Spillovers in Debt Markets," The Journal of Finance, forthcoming.

Goldstein, I., Jiang H., and Ng, D. T., 2017. "Investor flows and fragility in corporate bond funds," Journal of Financial Economics, vol. 126(3), pp. 592-613.

Morningstar, Inc. Morningstar Direct, http://corporate.morningstar.com/US/asp/subject.aspx?xmlfile=40.xml.

1. The views expressed here are strictly those of the authors and do not necessarily represent the views of the Federal Reserve Board or the Federal Reserve System. We thank Gene Kang for his excellent research assistance. Return to text

2. This data capture the population of loans with a committed balance of greater than or equal to \$1 million. Return to text

3. This is due to the inaccuracy in Y-14's obligor type classification based on North American Industry Classification System. Return to text

4. On the other hand, there are a few instances where other types of non-bank financial institutions such as private equity funds, REITs, and hedge funds are erroneously captured as mutual funds. In these cases, we use a comprehensive list of non-mutual fund/ETF keywords, such as "PRIVATE," various types of limited partnership entities, and well-known hedge fund names, to drop the obligors with such keywords in the names. Return to text

5. The N-CEN form was introduced as part of the SEC's efforts to modernize disclosures and enhance transparency by registered investment companies. The reporting is filed annually and provides new information on fund liquidity management via lines of credit, interfund lending/borrowing, etc. Return to text

6. General corporate purposes, permanent or short-term/seasonal working capital take up more than half of the observations. Return to text

7. On the demand (asset managers) side, while cash holdings remained relatively stable at around 5-6% during the same period, other types of liquidity management tools including interfund lending facilities, borrowing from affiliates, and use of derivatives could have contributed to the development as well. In addition, any supply (banks) side changes to credit extension or standards might have affected the reduction. Return to text

8. Such credit extension to mutual funds and ETFs is not economically significant from the lending bank's point of view either. For BHCs with the largest exposures to mutual funds and ETFs, the exposures in terms of their total assets are de minimis (less than 1%). However, because our data only capture bank loans coming from the largest banks reporting FR Y-14, the actual amount of bank credit obtained by mutual funds and ETFs would be larger than our estimates in this analysis. Return to text

9. Importantly, our sample is likely to serve as a conservative, lower-end estimate of the actual amounts of bank credit obtained by bank loan funds because credit lines are often shared by multiple funds under the same fund family. Return to text

10. Usage, or, the drawn amount, can only be defined for credit line facilities and therefore is reported only for credit lines and not for term loans. Return to text